As a data platform manager in a data-rich company I agree: programmers tend to prefer their favourite hammers to sql, even if a modern relational database can do the thing much better.
But! As somebody with a relatively good understanding of a history of sql, relational dbs and related concepts I have to add: SQL is often to blame.
All the the amazing engineering that goes into database engines, clean and coherent ideas of relational algebra, optimisability of a declarative approach to computaion - all of that gets bad rep because of the nightmare of sql-the-language.
The standard, the syntax, every little bit that could go wrong is just wrong from the point of view of a language designer. Composability, modularity, predictability, even core null-related defaults.
But it's everywhere. We just have to accept it.
But! As somebody with a relatively good understanding of a history of sql, relational dbs and related concepts I have to add: SQL is often to blame.
All the the amazing engineering that goes into database engines, clean and coherent ideas of relational algebra, optimisability of a declarative approach to computaion - all of that gets bad rep because of the nightmare of sql-the-language.
The standard, the syntax, every little bit that could go wrong is just wrong from the point of view of a language designer. Composability, modularity, predictability, even core null-related defaults.
But it's everywhere. We just have to accept it.